prikaz prve stranice dokumenta Robotska percepcija položaja ljudske ruke
  Download
PDF 1.76 MB
undergraduate thesis
Robotska percepcija položaja ljudske ruke
2023. urn:nbn:hr:200:382519
Grgić, Antea
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Cite this document

Grgić, A. (2023). Robotska percepcija položaja ljudske ruke (Undergraduate thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:382519

Grgić, Antea. "Robotska percepcija položaja ljudske ruke." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2023. https://urn.nsk.hr/urn:nbn:hr:200:382519

Grgić, Antea. "Robotska percepcija položaja ljudske ruke." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2023. https://urn.nsk.hr/urn:nbn:hr:200:382519

Grgić, A. (2023). 'Robotska percepcija položaja ljudske ruke', Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 20 April 2024, https://urn.nsk.hr/urn:nbn:hr:200:382519

Grgić A. Robotska percepcija položaja ljudske ruke [Undergraduate thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2023 [cited 2024 April 20] Available at: https://urn.nsk.hr/urn:nbn:hr:200:382519

A. Grgić, "Robotska percepcija položaja ljudske ruke", Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:200:382519

Please login to the repository to save this object to your list.
Metadata
TitleRobotska percepcija položaja ljudske ruke
Title (english)Human hand pose robot perception
AuthorAntea Grgić
MentorPetra Pejić (mentor)
MentorValentin Šimundić (sumentor)
Committee memberDamir Blažević (predsjednik povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Computer Engineering and Automation)
(Chair of Automation and Robotics)
Osijek
Defense date and country2023-09-22, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
U ovom završnom radu predstavljen je sustav u kojemu čovjek zadaje zadatke robotu simbolom prikazanog rukom. Nakon što je sakupljena baza RGB slika na kojima se nalazi ljudska ruka u nekoj od sljedećih poza: otvoreni dlan s ispruženim prstima, stisnuta šaka s podignutim palcem („like“) i stisnuta šaka s podignutim kažiprstom, slike su analizirane pomoću OpenPose modela. Pomoću modela, na slikama je detektirana ljudska ruka i značajke koje predstavljaju set od 21 točke zglobova prstiju i šake. Potom, izvađena polja točaka spremljena su u CSV datoteku i transformirana u novi koordinatni sustav s ishodištem u zglobu zapešća. Takvi podatci dani su klasifikatoru Random Forests kao ulaz. Na temelju međusobnog položaja transformiranih značajki treniran je klasifikator. Skup za treniranje i testiranje raspoređen je u omjeru 80:20. Uspješnost klasifikatora izmjerena je preciznošću i točnošću od 98.5%. Eksperimenti su provedeni i na stvarnom robotu UR5 pomoću ROS-a, gdje je na temelju izlaza iz klasifikatora, robot izvršavao predefinirane kretnje.
Abstract (english)
In this final paper, a system is presented in which a person assigns tasks to a robot with the symbol shown by the hand. After collecting a database of RGB images of a human hand in one of the following poses: open palm with fingers extended, clenched fist with thumb raised ("like"), and clenched fist with index finger raised, the images were analyzed using the OpenPose model. Using the model, a human hand and features representing a set of 21 points of the finger and hand joints are detected in the images. Then, the extracted point fields were saved in a CSV file and transformed into a new coordinate system with the origin at the wrist. This data is given as input to the Random Forests classifier. The classifier was trained based on the mutual position of the transformed features. The set for training and testing is distributed in a ratio of 80:20. The success of the classifier was measured with a precision and accuracy of 98.5%. Experiments were also conducted with the real UR5 robot using ROS, where the robot performed predefined motions based on the output of the classifier.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:382519
Study programmeTitle: Undergraduate University Study Programme in Computer Engineering
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2023-10-18 09:46:44